Data communications between computing devices is established using a data network. A data network includes network interfaces installed at/in each of the computing devices (“nodes”) and electrical or optical physical media connecting those network interfaces to each other. Communication information is transformed into electrical or optical signals and then transported between the nodes using the physical media. In order to realize a communication network consisting of multiple nodes (sometimes referred to as a multi-node-network), network aggregation devices (NADs) in the form of hubs, switches or routers are often used to route the flow of data from the data's originating node to the data's destination node.
In some configurations, a NAD is connected to all nodes of the network, or a subset of the nodes of the network, either through a direct physical connection (or a link), or an indirect physical connection through other NADs. A full connection between the nodes (e.g. an all-to-all connection) may be prohibitively complex for practical purposes.
In some configurations, a network having many nodes is realized by defining hierarchical layers of NADs, whereby nodes on the same level are connected and communicate with the help of routing nodes/NADs of the higher hierarchical level. In an optical communication network, the data communication between the NADs (e.g. using the physical medium) is often conveyed by optical waves, whereas the NADs manipulate the electrical embodiment of signals/data. Thus, the NADs often include electrical-to-optical, optical-to-electrical, and optical-to-electrical-to-optical signal conversion modules. These modules can negatively impact energy efficiency, and complexity, of the NAD. Further, NADs include a finite number of ports to which physical connections to nodes and other NADs are made. Thus, NAD data traffic handling capacity is often bounded by the internal processing and the network interface throughput limit of each NAD. As such, conventional data network architectures are challenged by fundamental limitations in energy efficiency as well as scalability.